Artificial intelligence is changing the way teams work, and OpenAI just took another concrete step in that direction.
The company announced Workspace Agents in ChatGPT, a new feature that promises to transform how teams handle complex tasks and workflows that demand heavy coordination.
We are not talking about a simple assistant that answers questions.
These agents run in the cloud, keep working even when you are offline, and can be shared across your entire team. They are powered by Codex, which gives them access to a full environment with files, code, tools, and memory.
Imagine having a digital coworker that never stops, follows your company processes, asks for approval when needed, and even integrates with Slack.
That is exactly what OpenAI is bringing to the table.
In this article, you will learn what these agents are, how they work in practice, which use cases are already running inside OpenAI itself, what controls are available for admins, and what to expect from them over the coming months. 🚀
What Are Workspace Agents in ChatGPT
Workspace Agents are a direct evolution of the GPTs that already existed inside ChatGPT. The fundamental difference is that while GPTs were basically customized assistants for responding to prompts, the new agents are designed to execute entire sequences of tasks autonomously, without you needing to monitor every step of the process. They receive a goal, understand your business context, and carry out the necessary actions to reach the expected result.
Powered by Codex in the cloud, these agents go beyond simply generating text. They can write and run code, use connected apps, remember what they learned from previous interactions, and continue working across multiple steps. It is a paradigm shift from what most people were used to doing with generative artificial intelligence.
The big difference here is persistence. Unlike a regular conversation in ChatGPT, where everything happens in real time and depends on you being present, Workspace Agents run in the cloud continuously. That means you can set up an agent to process reports, classify data, respond to internal requests, or even prepare weekly summaries, and it will handle all of that while you are in another meeting, sleeping, or simply focused on something more strategic. The idea is that operational work never stalls because of a lack of human attention.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Another point worth highlighting is the ability to share agents across teams within the organization. An agent configured by one team member can be used by the entire group, either directly in ChatGPT or through Slack, maintaining consistency in processes and ensuring everyone follows the same workflow standard. And since agents have memory and can be guided and corrected through conversation, they improve as they are used. Over time, they become a practical way to keep team knowledge up to date: build once, improve with use, and share or duplicate for new workflows.
How to Create a Workspace Agent in Practice
Creating a Workspace Agent is surprisingly simple. Just click on Agents in the ChatGPT sidebar and describe a workflow your team handles frequently. From there, ChatGPT itself walks you through it step by step to turn that description into a functional agent. It defines the steps, connects the right tools, adds capabilities, and tests everything until the agent works as expected.
You can also just drop in a file and let ChatGPT interpret the need and build the agent from that context. On top of that, there are ready-made templates for areas like finance, sales, marketing, and more. Each template comes with built-in capabilities and suggested tools, which speeds up the initial setup significantly and lets you customize from a solid foundation.
Workspace Agents are available in research preview for ChatGPT Business, Enterprise, Edu, and Teachers plans. An important note: traditional GPTs remain available while teams test the new agents, and OpenAI will soon make it easier to convert existing GPTs into Workspace Agents.
Real Use Cases Inside OpenAI
OpenAI is not just selling an idea. The company’s internal teams are already using Workspace Agents day to day, and the examples are pretty revealing when it comes to the potential of this tool.
Here are some of the agents that OpenAI teams have built — and that any team can replicate:
- Software Reviewer: Analyzes software requests from employees, checks whether they align with approved tools and internal policies, recommends next steps, and opens IT tickets when necessary.
- Product Feedback Router: Monitors Slack channels, support channels, and public forums, turns feedback into prioritized tickets, and generates weekly product summaries.
- Weekly Metrics Reporter: Pulls data every Friday, creates charts, writes the summary, and shares the report with the team automatically.
- Lead Outreach Agent: Researches inbound leads, scores each one based on the team’s qualification criteria, drafts personalized follow-up emails, and updates the CRM.
- Third-Party Risk Manager: Researches vendors, evaluates signals like sanctions exposure, financial health, and reputational risk, and produces a structured report.
OpenAI’s sales team, for example, uses an agent that gathers details from call notes and account research, qualifies new leads, and drafts follow-up emails directly in the rep’s inbox. This helps account teams spend less time piecing together scattered information and more time talking to customers.
The accounting team built an agent that prepares essential parts of the monthly close, from journal entries to balance sheet reconciliations and variance analyses. It completes the work in minutes, generates workpapers with the necessary inputs and control totals for review, and follows the company’s internal policies. The agent is available in ChatGPT for anyone on the team to use, or added to Slack channels so the team can ask questions and collaborate around the results.
Slack Integration and Continuous Automation
The integration with Slack is one of the most interesting features of Workspace Agents. The agents can be deployed directly into Slack channels, where they can pick up requests as they come in, send updates, ask for confirmations, and even receive instructions — all within the communication channel the team already uses every day.
On top of that, you can configure agents to run on scheduled intervals. That weekly report someone always forgets to prepare? An agent can handle it automatically every Friday. That ticket triage that eats up two hours a day? An agent can take care of it in the background, without interruption.
OpenAI’s product team, for example, built an agent that proactively answers employee questions in Slack channels. The agent provides a clear response, links relevant documentation, and opens a ticket when it identifies a new issue. This helps teams unblock faster while making sure important follow-ups don’t slip through the cracks.
Today, teams can interact with agents in ChatGPT and Slack, with more surfaces coming soon. The idea is for agents to participate in the conversations and workflows where work already happens, helping move things forward with less manual coordination.
Team Collaboration Reimagined with AI
Team collaboration has always been one of the biggest challenges in companies, especially when teams grow, are distributed across different time zones, or need to coordinate work involving lots of dependencies between departments. Artificial intelligence has already helped people work faster individually, but many of the most important workflows inside an organization depend on shared context, handoffs, and decisions across teams. Workspace Agents were designed for exactly this kind of work.
They can pull context from the right systems, follow team processes, request approval when needed, and keep work flowing across different tools. This eliminates that classic scenario where an email gets lost in an inbox, an approval stalls because the responsible person was in a meeting, or a report goes unfinished because it depended on data sitting in another system.
Knowledge inside companies often ends up scattered across people and systems. Workspace Agents offer a way to turn that scattered knowledge into a reusable workflow: one that follows the right process, uses the right tools, and can be shared across the entire organization. Build once, improve with use, and replicate for new needs.
The most transformative aspect, however, is how these agents change the relationship teams have with operational work. When repetitive and coordination-heavy tasks are absorbed by intelligent systems, professionals gain more space to think strategically, create, and solve problems that actually require human judgment. Artificial intelligence does not replace the team — it amplifies each person’s capacity, allowing the group as a whole to deliver more with the same resources.
Controls, Security, and Monitoring for Admins
When you delegate work to an agent, control stays with the human. You decide which tools and data the agent can use, which actions it can take, and when it needs approval. For sensitive steps — like editing a spreadsheet, sending an email, or adding a calendar event — you can require the agent to ask for permission before moving forward.
After sharing an agent, built-in analytics show how it is being used, including how many runs have been completed and how many people are using it. This gives you visibility into the agent’s real impact on the team.
Workspace Agents come with enterprise-grade monitoring and controls, allowing admins to protect sensitive data while giving teams a secure way to accelerate with AI. ChatGPT Enterprise and Edu admins can control which connected tools and actions each user group can access, as well as manage who has permission to use, build, and share agents.
Built-in safeguards help agents stay aligned with the instructions they were given when they encounter deceptive external content, including prompt injection attacks. The Compliance API gives admins visibility into the configuration, updates, and runs of each agent, allowing them to monitor and control how agents are being built and used. Admins can also suspend agents when necessary.
Soon, admins will also be able to view all agents built within their organization directly in the admin console, including usage patterns and connected data sources.
What Companies Are Already Saying
Early testers of Workspace Agents are already reporting more consistent results and more time freed up for higher-value work. Ankur Bhatt, from AI Engineering at Rippling, shared that the hard part of building an agent is not the model itself, but the integrations, memory, and user experience. According to him, Workspace Agents condensed all of that work, allowing a sales consultant to build, evaluate, and iterate on a sales opportunity agent from start to finish — without needing an engineering team. The agent researches accounts, summarizes calls from Gong, and posts deal summaries directly in the team’s Slack channel. What used to take reps 5 to 6 hours per week now runs automatically in the background for every deal.
Pricing and Availability
Workspace Agents are available in research preview for ChatGPT Business, Enterprise, Edu, and Teachers plans. For Enterprise and Edu plans, admins can enable agents using role-based controls.
A relevant note on costs: Workspace Agents will be free until May 6, 2026, when a credit-based pricing model will take effect. That gives teams plenty of time to experiment, test, and understand the value of agents before any financial commitment.
What to Expect in the Coming Months
OpenAI has already signaled that it will keep adding improvements over the coming weeks to help teams do more with less manual effort. Among the planned updates are:
- New triggers that can kick off work automatically
- Enhanced dashboards to understand and improve agent performance
- More action types for agents to interact with business tools
- Support for Workspace Agents in the Codex app
- Easier conversion of existing GPTs into Workspace Agents
The market for task automation with artificial intelligence is growing fast, and Workspace Agents represent a concrete bet by OpenAI to position itself as the central work platform for teams that want to use AI for real — not just for one-off queries, but as an active part of their daily processes. 🤖
Summary of Key Workspace Agent Capabilities
- Powered by Codex in the cloud with access to files, code, tools, and memory
- Autonomous task execution even when the user is offline
- Cross-team sharing with standardized and reusable configurations
- Native integration with ChatGPT and Slack for seamless communication
- Scheduled runs for automatic execution of recurring tasks
- Human approval requests at critical workflow steps
- Persistent memory that improves with use over time
- Built-in analytics to track usage and results
- Enterprise controls with Compliance API and prompt injection protection
- Ready-made templates for finance, sales, marketing, and more
Teams do their best work when knowledge is easier to find, processes are easier to follow, and people can get help in the flow of work itself. Workspace Agents are an early step in that direction: artificial intelligence that works alongside people in the tools and conversations where work already happens, helping teams spend less time coordinating and more time creating, building, and making decisions that move the business forward. 💡
